Works with any AI tool that supports MCPWorks with any MCP clientBring AI visibility
Bring AI visibility
into your AI
Let the AI you already use read your brand share, raw AI answers, and the sources they cite.
LV
Category citation sourcesShare
When ChatGPT recommends air purifiers, which sites does it cite most? Which should we go after first?
Called GEOly · get_public_categorySites ChatGPT cites most · last 30 days
Answers in this category cited 367 sites in total. The top ones:
1
rtings.comReviews25.9%
rtings.comReviews25.9%2
epa.govGov11.9%
epa.govGov11.9%3
reddit.comForum6.1%
reddit.comForum6.1%4
cleanaircritic.comReviews3.5%
cleanaircritic.comReviews3.5%5
consumerreports.orgReviews2.7%
consumerreports.orgReviews2.7%rtings.com alone accounts for a quarter, and the top three add up to 43.9%. Check whether rtings.com reviews your models first, then look at the Reddit threads.
Reply…
Sample data: air purifiers · US · ChatGPT · Aug 6 – Sep 4, 2026, GEOly public category snapshot
Works in the AI you already use.Claude, ChatGPT, Cursor, and any AI tool that supports MCP.
No switching back to a dashboard.The data shows up in the chat you’re already in.
Ask in plain words
No dashboards, no filters, no tool names to remember. The AI picks the right tool.
Your brand, plus the whole category
Visibility, raw answers, cited sites, competitors, and public category rankings.
Read-only by default, revoke anytime
Choose read or write per data type. With write access, the AI can add prompts and competitors for you.
Share, quotes, rivals.Under each card is a question you can ask as-is.
Brand ranking · Air purifiersMention share
LevoitYou17.9%
Blueair17.6%
Coway16.0%
Winix9.9%
IQAir3.8%Check your rank“Where do we rank in this category, and who’s right behind us?”
ChatGPT answer08/21
Which bedroom air purifiers are the quietest at night?
Best overall: Levoit Core 300S. It's exceptionally bedroom-friendly: quiet, compact, has a dedicated Sleep mode that shuts off the display, and has an air-quality sensor.
Read the quotes“When AI recommends us, what exactly does it say?”
Topics where you’re missingMention share
Car air purifiers
You 0%IQAir 25.2%
Humidifier-purifier combos
You 1.3%Dyson 23.3%
Find the gaps“In which topics does AI only recommend our competitors?”
55 tools, one catalog.What’s free, what costs credits, what needs write access.
Endpoint https://app.geoly.ai/api/mcp/v1. How many a given connection sees depends on plan, single vs multi-brand, and whether you granted write access: up to 55 on a single-org, multi-brand, write-enabled Grow-tier connection; 48 on a read-only multi-org one.
ADiscovery & routing6 toolsFree
get_brand_contextCall this first, once per run. One shot: brand profile, org plan, GA4 / Cloudflare integration status, the three date axes, first-run progress, entitled platforms, topics, filter value sets, tracked competitors, and credits left.list_brandsList the brands in the org to get a brand_id (multi-brand and multi-org connections only).list_organizationsList the orgs this connection can reach (multi-org only).get_quotaCredits left this month and the billing period. -1 means unlimited, not overdrawn.get_current_dateServer date; in a normal flow get_brand_context.today already covers it.resolve_page_contextTurn the GEOly page URL you are looking at into an entity scope — page awareness for desktop clients.BBrand-own · overview & KPI3 toolsCosts credits per call
get_brand_overviewThe one headline KPI call, value for value with the /overview KPI row: visibility (AIGVR), mention rate, citation rate, sentiment, and brand share (the shareOfVoice field). Add include_win_rate for first-position rate.get_daily_trendDay-by-day KPI series (= the /overview trend chart). breakdown takes total / platform / topic; compare_previous returns the prior window alongside.resolve_my_brand_publicBridge your monitored brand to its public category brand, so you can chain into the group J industry tools. Free.CBrand-own · prompts & answers8 toolsCosts credits per call
get_prompt_listPrompt performance table in three views: table (= the /prompts table), mention_rates (per-prompt mention rate ascending — the one for intent blind spots), and catalog. Filter branded vs unbranded.get_prompt_detailOne prompt, windowed (30 days by default) and scopable to a platform (= /prompts/[id]).list_prompt_recordsOne prompt's full execution history: every daily record on every platform, paginated — not just the latest.list_brand_answersBrand-level answers table (= /answers): every completed answer across all prompts in the window, newest first, filterable by platform and topic.get_prompt_record_detailOne record in full: the raw answer, its citations, sentiment, and shopping cards — the same read model as the in-app answer dialog.get_prompt_citationsOne prompt's citations: raw per-record rows, or a deduplicated URL list with shares.get_brand_search_queriesQuery fanout — the real web searches ChatGPT issued while answering your tracked prompts.get_topic_listMonitored topics (= the /prompts left rail), with active and archived counts. Free.DBrand-own · citations, domains & pages3 toolsCosts credits per call
get_citation_overviewThe /sources/citations domains tab in three sections: board (the domain leaderboard), table (rows; gap_only=true goes straight to the gaps where competitors are cited and you are not), and links (outbound links inside the answer text).get_domain_detailOne domain's citation profile: citations, share, whether it also mentions you, and which prompts it shows up in.get_url_detailOne page URL, in two window calibers (rolling / page). The two numbers are not interchangeable — quote the caliber you asked for.EBrand-own · competitors, platforms, sentiment, shelf8 toolsCosts credits per call
get_brand_boardTracked-brand board (caliber=tracked_brand_v1), value for value with the /overview board. It counts only the brands you track — a brand you never added never appears in the shares.get_platform_matrixYour brand plus automatically discovered competitors × platform, by competitor or by topic.get_competitor_cooccurrenceWhich competitors get named in the same answer as you: per-record detail plus a summary. The data behind competitor-narrative work.get_competitor_listThe brand library (= Settings › Brand › Brands): tracked, suggested, removed.get_verdictThe AI Verdict page (/perception), one board per view: competitors (who beats you), sources, and more.get_topic_analyticsPer-topic answer analytics: sentiment distribution, response types, competitor co-occurrence — the topic detail page read model.get_shopping_performanceAI shelf (= /shopping, the ChatGPT and Google AI Mode shopping cards): trigger rate, your appearance rate and shelf share, brand board, seller board, and the prompts where your products are absent.get_sentiment_dashboardRecord-level sentiment labels. Not the /perception caliber — for that use get_verdict.FBrand-own · site audits & AI traffic3 toolsFree
get_audit_listGEO site audit history (= /audit), both whole-site and single-page runs.get_audit_detailOne audit report: fail / warning / pass items, with per-page detail.get_traffic_data/my-website/ai-traffic: source=ga4 or cloudflare — the real visits arriving from AI.GBrand-own · writes8 toolsNeeds write access at authorization
create_promptCreate a monitoring prompt: text, category, priority, country (US by default), topic, tags.archive_promptArchive a prompt (stops monitoring, cancels in-flight collection); restore=true brings it back. Idempotent.update_prompt_tagsBulk add / remove tags (≤500 prompts), or rename one tag brand-wide.move_prompts_to_topicMove prompts into a topic, or topic_id=null to ungroup them.create_topicCreate a prompt topic (unique within the brand, case-insensitive).create_brandOrganization-level: add a new monitored brand, the same path as the in-app "Add brand" dialog. Plan brand limits and your role apply.create_competitorAdd a brand to the brand library; a previously removed brand gets re-tracked.trigger_promptRun monitoring for a prompt right now. Consumes monitoring quota.HReports1 toolFree · no brand needed
get_agent_ready_scansAgent Readiness scan history; pass scan_id for one scan’s result.IPublic source domains3 toolsFree · on every connection
get_public_sources_overviewThe most-cited source domains across all topics. Each row’s score is AI DA (0–100 AI domain authority).get_public_source_domain_detailProfile one citation source domain, optionally with its scorecard.get_public_source_brand_conduitThe top 30 topics where one domain’s citations co-occur with one brand — the contexts in which that source funnels AI toward it.JPublic industry data13 toolsGrow and above · mostly priced per row
search_public_entitiesName or domain → public ids and slugs. Call this first, free.list_public_topicsBrowse and paginate public topics. Free.get_public_coverageThree views: locales, platforms (valid platforms for a scope), and data_window (the time anchor — get it before quoting numbers). Free.get_public_topicTopic layer: overview, brand_leaderboard, som_trend, prompt_matrix, prompts, citation_domains, difficulty.get_public_topic_prompt_detailOne public prompt over all history; include_fanout adds the query fan-out block.get_public_topic_record_detailOne public AI answer, as a record sheet.get_public_brandPublic brand profile. Pass brand_ids for a multi-brand comparison; views include visibility, footprint, competitors, citations, category_ranking, shopping, and rank_citation (AI Overview rank × citations).get_public_categoryCategory overview: leaderboard, SoM trend, topics, citation domains, recent mentions.get_category_whitespaceCategory whitespace: topics no brand owns yet.get_category_brand_momentumBrand momentum: who is rising and who is falling.get_public_search_queriesAI search demand: territories, product_spaces (free), query_detail, theme_detail.list_public_shopping_productsAI shelf rankings: products (one category’s shelf) and boards (across categories).get_public_shopping_product_detailOne product in full, or a cheap card preview.How it’s billed
Credits are one org-wide monthly pool — get_quota shows what is left, and -1 means unlimited. Brand-own monitoring reads are priced per call (1 / 3 / 10 credits regardless of page size: get_daily_trend 1, get_brand_overview 3, get_citation_overview 10 for board and 3 for table, get_platform_matrix / get_topic_analytics / get_brand_search_queries 10). Most public industry views are priced per row returned. Site audits, AI traffic, Agent Readiness, the write tools, and every discovery or navigation tool are free.
Connected in a minute.Nothing to install, no keys to copy.
01Add the URL
Pick your client below and copy its config.
02Authorize in the browser
Sign in to GEOly, pick an organization, choose read or write per data type. Read-only by default.
03Just ask
Go back to your chat and ask in plain words.
Run in your terminal
claude mcp add --transport http geoly "https://app.geoly.ai/api/mcp/v1"Then run claude /mcp, pick geoly, and authorize in the browser.
You can also query data from your terminal and scripts with the GEOly CLI, or install the Agent Skill — it ships seven slash commands, each returning a finished report.
FAQ
What people ask most before connecting
Any AI tool that supports remote MCP (Streamable HTTP) and OAuth. Add the URL and authorize in the browser. Claude Code, Cursor, Claude Desktop, VS Code, Windsurf, and Codex have ready-made configs.
Connect GEOly MCP today
Free 3-day trial. Connect with one URL and authorize in your browser.
GEOly MCP use cases
Each use case is a question you can paste into any AI tool connected to GEOly MCP, the tools it calls, what comes back, and what to do with it.
- Brand & marketing teamsWeekly Share of Model report in your AI chatSomeone exports charts from the dashboard every Monday and rewrites the same summary. The numbers are already in GEOly; the writing is what takes the time.
- SEO & content teamsFind topics where AI only recommends your competitorsOverall visibility can look fine while whole sub-topics go to competitors. Those gaps are where new content and outreach pay off first.
- PR & outreach teamsBuild an outreach list from the sites AI citesOutreach lists are usually built from backlink or traffic data. AI answers lean on a different set of review sites, forums and retailers, and some high-traffic sites are never cited at all.
- Brand & product teamsCheck exactly what AI says about your brandA score tells you how often you appear, not what is being said. Wrong prices, discontinued products or a mix-up with another brand only show up in the answer text.
- GEO & SEO agenciesClient-ready AI visibility summary for agenciesAgencies report to clients in their own format. Copying numbers out of a dashboard into a deck for every client every month does not scale.
- Strategy, sales & agenciesResearch a category's AI landscape before a pitchBefore a pitch or a market entry you need to know who AI recommends in a category, but you have no tracked prompts there yet.
Built something useful with GEOly MCP? Submit it to the MCP Challenge to be featured as a Community Built use case.
Join the MCP Challenge